Four bite-size games to master p-values, one- vs two-sided tests, multiplicity, and the “N boosts significance” effect.
Single file • Mobile-friendly
1) Decision Clinic
Look at the study summary and choose the best next step. Think: p-value + confidence interval + MCID + context.
Score
Correct:
0
Attempts:
0
Tip: “Significant” can still be clinically trivial if the effect is below the MCID. “Not significant” can still be promising if the CI includes meaningful benefit.
2) One-sided or Two-sided?
Read the question. Choose whether the appropriate test is one-sided or two-sided — assuming it’s pre-specified.
Score: 0/0
3) Multiplicity Challenge
You tested multiple outcomes. Enter the unadjusted p and the number of tests. Decide if it’s still significant after Bonferroni.
Adjusted p (Bonferroni)
—
Rule: padj = min(1, p × m). Bonferroni is conservative but easy to remember.
4) P-Value Builder
See how sample size (N) and effect size change the p-value. Model: two-sample difference in means, SD=1 each arm, equal N.
Computation (normal approx)
—
SE ≈ √(2/n) when SD=1 in each arm. z = effect / SE. Two-sided p = 2×(1−Φ(|z|)).